CHAPTER 4: THE ROLE OF ATTACHMENT ON PSYCHOPATHIC TRAITS
4.2.4. Attachment and the Five Factor Model (FFM) of personality
There have also been a number of studies investigating whether FFM personality traits are related to attachment. Such studies show that higher levels of Extraversion (E) are
associated with greater attachment security to parents and this link has been shown in cross-section studies with children (Hagekull & Bohlin, 2003), adults (Roisman et al., 2007) as well as in a longitudinal study from childhood to adulthood (Fransson, Granqvist, Bohlin, & Hagekull, 2013). As mentioned previously, secure attachment means that an individual has a positive image of the self and others. They feel that they are deserving of love and assume that others will accept them. Therefore these individuals will be more likely to socialise with others and be more outgoing (high E) than individuals who feel that others will not like and accept them (Fransson et al., 2013).
121 Neuroticism (N) has also been related to attachment security (Hagekull & Bohlin, 2003; Fransson et al., 2012). This link showing higher N being associated with more attachment insecurity may exist because attachment is implicated in the development of emotion regulation and coping strategies and these aspects are associated negatively with N.
Some research also suggests a link between insecure attachment and low Conscientiousness (C) perhaps because C is to some extent related to cognitive processes, such as attention and executive function which in turn have been found to be associated with attachment (see Fransson et al., 2013).
Lastly, it has been found that high Openness (O) is linked to attachment security, such that those with secure attachment styles are higher in O. Arguably this is because securely attached children are allowed to, and feel safe to, explore their surroundings (Fransson et al., 2013; Hagekull & Bohlin, 2003).
The evidence presented here demonstrates that FFM personality traits are, to some level, involved with attachment and may also influence the relationship between attachment and psychopathy.
4.2.5. Current study
Attachment could be a potential underlying mechanism that is involved in the emotional deficits seen in psychopathic individuals. To investigate this claim it needs to be established first whether a link between attachment and psychopathic traits can be found in a non- forensic population. This current study investigates undergraduate students and assesses attachment for both parents as well as attachment styles. Psychopathic traits are measured using the Youth Psychopathy Inventory (YPI: Andershed et al., 2002). We also examine if the relationship between attachment and psychopathic traits is moderated by standard
122 personality traits to assess if these traits influence the relationship between attachment and psychopathy. To our knowledge this is the first time this has been examined.
Attachment interviews and self-report scales have been developed to tap into those bonds with attachment figures as well as the bonds individuals have in relationships or friendships. The Inventory of Parent and Peer Attachment revised (IPPAr: Armsden & Greenberg, 1987) is a self-report measure that is used here to assess attachment behaviours and feelings in relation to important attachment figures. Additionally, the Relationship Questionnaire (RQ: Bartholomew & Horowitz, 1991) is a very short attachment style questionnaire that gives descriptions of the four different attachment styles discussed earlier in this chapter (secure, avoidant, preoccupied and dismissing-avoidant). These attachment questionnaires are chosen for this study in order to capture a broad range of important factors, i.e. not only mother attachment but also attachment to fathers, and not only global attachment, but each of the different attachment styles. However, due to finding that peer attachment does not seem to be associated with psychopathic traits (Kosson, Cyterski, et al., 2002; Marcus & Betzer, 1996) and the issue of high drop-out rates in online surveys, it was decided to exclude the peer attachment scale in this study.
It is hypothesised that psychopathic traits will be negatively associated with attachment security to mothers and fathers. Since the literature is inconsistent as to how attachment to parents is related to the different psychopathy factors, no specific predictions were made regarding whether this relationship would be found for primary or secondary psychopathy specifically. It is hypothesised that high E and high A would be linked to secure attachment whereas high N would be linked to more insecure attachment. These FFM traits are also examined to see if they moderate the relationship between attachment and psychopathy. As
123 O has generally been found to be unrelated to psychopathy no moderation effect was
expected. However, it is possible that O is directly related to attachment because securely attached children are free and encouraged to explore their environment and this curiosity could be linked to O (Fransson et al., 2013).
Since some of the research presented here suggests an inverse relationship between psychopathy and attachment security it is expected that the YPI higher-order factors will be negatively related to the secure attachment style.
Given that dismissing-avoidant attachment is linked to a positive self-image and a negative image of others it is expected that this style will be related particularly to primary
psychopathic traits (i.e. Interpersonal and Callous/Unemotional YPI factors). Dismissing- avoidant attachment is also expected to be related to low A and low C because this
attachment style is most strongly related to antisocial tendencies (Main & Solomon, 1986). This has been demonstrated with the example of aggression in the previous chapters. The avoidant attachment style is related to a negative image of the self and others and is therefore thought to be more strongly related to secondary psychopathy, in line with Mack and colleagues findings (Mack et al., 2011). Preoccupied attachment is not expected to be related with psychopathic traits as this attachment style is related to seeking very close contact and approval in relationships which is not typically found in the psychopathy
literature. Avoidant and preoccupied attachment styles are expected to be related to high N as they are related to anxiety. Avoidant attachment should also be related to low E and low A because such individuals tend to avoid interpersonal relationships.
In summary, the YPI psychopathy factors are hypothesised to be negatively associated with attachment security to both mothers and fathers. High E, high A and high O are
124 hypothesised to be linked to secure attachment. N on the other hand is expected to be negatively related to secure attachment. It is also investigated whether these FFM traits, apart from O, moderate the effect between attachment and psychopathy.
For attachment styles it is expected that the YPI factors will be negatively associated with secure attachment. Dismissing-avoidant attachment is hypothesised to be linked particularly to higher YPI Interpersonal and CU traits. Additionally, it is hypothesised that the avoidant attachment style will be related to higher YPI Lifestyle traits. No relationship is expected between psychopathic traits and the preoccupied attachment style. For FFM traits it is hypothesised that the preoccupied and avoidant attachment style will be related to high N with additional associations for avoidant attachment and low E and low A. Finally, the dismissing-avoidant attachment style is expected to be associated with low A and low C.
4.3. METHOD
4.3.1. Participants
Undergraduate students (n= 338) were recruited from the University of Nottingham. They were contacted via email through the departments of the University of Nottingham and the Personality Social Psychology and Health (PSPH) group participant pool. Participants were included if they were aged between 17-21 years and had completed all assessments. 109 participants were excluded due to them not completing all measures; leaving a total of 229 participants (59 males, 170 females, M= 19.24 years, SD= .89).
4.3.2. Measures
125
Youth Psychopathy Inventory see Chapter 3 (section 3.2.1.2. Measures).
Inventory of Parent and Peer Attachment-revised (IPPAr: Armsden & Greenberg, 2009) is a
self-report measure that was used to assess attachment behaviours and feelings in relation to important attachment figures (three sub-scales: attachment to mother, father and close peers). However, in this study the peer attachment scale was not included as discussed in section 4.2.5. The IPPAr therefore included 25 items for each attachment figure, which produces two different attachment scores; one for attachment to mothers and one for attachment to fathers (see Appendix 4.2.). Items assess trust, communication and alienation to parents and were measured on a 5-point Likert scale from “Almost never true or never
true” (1) to “Almost always or always true” (5). The scale has previously shown good internal
reliability for mother attachment (α=.87), and father attachment (α=.89) (Armsden & Greenberg, 2009).
Relationship Questionnaire (RQ: Bartholomew & Horowitz, 1991) is a short attachment style
questionnaire. It gives short descriptions of four different attachment styles (secure, fearful- avoidant, preoccupied and dismissing-avoidant) from which the one that describes the individual best should be chosen. Each participant also rates each attachment style on a 7- point Likert scale from “Disagree strongly” (1) to “Agree strongly” (7) in relation to how well or poorly the description fits the individual (see Appendix 4.3.).
4.3.3. Procedure
Participants were contacted via email through the University of Nottingham academic departments and the PSPH mailing list. They were invited to take part in an online survey accessible through a link given in the contact email. First participants were presented with an information sheet and consent form (see Appendix 4.1.). Once they completed the
126 consent form they filled in the four questionnaires in the order described in the previous section (4.3.2.).
For taking part in the survey participants were entered into a £25 prize draw if they wished. Participants’ responses and contact details were kept separately in order to maintain anonymity. This was done by instructing participants at the end of the study to send a separate email to the researcher if they wanted to be entered into the prize draw. This ensured that email addresses were saved in a separate file from participants’ responses. One person was then selected by means of a random number generator and received a £25 voucher via email.
4.3.4. Data analysis
Attachment scores on the IPPAr were summed separately for attachment to mothers and fathers to obtain a total attachment score for each parent. A higher score represented more secure attachment. For the YPI, average scores for all 10 sub-factors as well as the three higher-order factors (Interpersonal, CU and Lifestyle) were calculated.
Attachment and YPI scores were checked for normality. Attachment scores for both mothers and fathers as well as ratings on the RQ for the different attachment styles were not
normally distributed. Subsequent analyses for these measures therefore used non- parametric tests where appropriate.
4.4. RESULTS
This study aimed to examine the relationship between psychopathic traits and attachment. Initially, correlations were conducted between parent attachment and YPI higher-order
127 factors (Interpersonal, CU and Lifestyle) and FFM personality traits (see Table 4.1). More insecure attachment to mothers and fathers was significantly correlated with higher levels of psychopathic traits. Amongst the FFM personality traits, Openness was the only trait that was not significantly correlated with attachment to parents and YPI higher-order factors. Therefore O was no longer included in any follow-up analyses. N was negatively correlated with secure mother and father attachment whereas E and A were positively correlated with secure attachment to parents in line with our hypothesis. C was positively correlated only with secure mother attachment.
To examine the ability of attachment to parents to predict psychopathic traits, regression analyses were conducted. Mother and father attachment variables were entered
simultaneously into the model as predictors and the psychopathic factors YPI Interpersonal, YPI CU and YPI Lifestyle were used as outcome variables in 3 separate analyses. Table 4.2 below summarises the results. Attachment to mothers was predictive of all YPI factors showing that less secure attachment was related to higher psychopathy scores. This relationship was strongest for YPI Lifestyle. However, R2 in all three models was low with parental attachment explaining only 9% of the variance in YPI Lifestyle and only 6% and 8% of the variance in the YPI Interpersonal and CU factors, respectively.
Next it was investigated whether FFM traits moderated the relationship between
attachment and psychopathic traits (Table 4.3-Table 4.5). Separate hierarchical regression analyses were conducted for YPI Interpersonal, YPI CU and YPI Lifestyle as outcome variables. The relationship of mother and father attachment with the YPI factors was investigated in separate analyses due to the attachment variables correlating with different numbers of FFM traits. For each analysis one FFM trait was used as moderator. The other traits were
128 entered as covariates in step 1. The interaction term was entered in step 3 leaving step 2 with the attachment variable and the moderator. These analyses were conducted to maintain an acceptable variable-to-sample size ratio. All predictor variables were mean- centred. The analyses showed that with the FFM traits added to the model, mother
attachment remained a significant predictor of all YPI higher-order factors (ps<.05). Father attachment was no longer significantly related to the Interpersonal or CU factor (p>.05). However, father attachment significantly predicted scores on the Lifestyle factor (p<.05). This relationship was moderated by N (see Figure 4.1).
A subsequent simple slope analysis was conducted with 1 standard deviation (SD) above and below the moderator variable (N) as suggested by Preacher (Preacher, Curran & Bauer, 2006). This showed that the relationship between father attachment and the YPI Lifestyle factor was strongest for individuals with moderate to higher levels of N. The analysis showed that insecure attachment to fathers seemed to remove the protective effect of being high on N, therefore increasing psychopathic traits (i.e. the Lifestyle factor).
129 Table 4.1 Spearman's rho correlation coefficients for YPI factors, FFM personality traits and IPPA attachment
Factors α 1 2 3 4 5 6 7 8 9 1. YPI Interpersonal .92 - 2. YPI CU .88 .64** - 3. YPI Lifestyle .84 .50** .42** - 4. IPPAr mother .95 -.26** -.28** -.29** - 5. IPPAr father .95 -.22** -.20** -.23** .55** - 6. O .63 -.01 -.12 .09 .06 .03 - 7. C .86 .17** -.15* -.45** .20** .12 -.05 - 8. E .77 .10 -.08 .20** .18** .24** .08 .15* - 9. A .79 -.51** -.57** -.35** .31** .35** .11 .20** .21** - 10. N .88 -.06 -.12 .01 -.21** -.27** .01 -.31** -.40** -.21**
Note. YPI= Youth Psychopathy Inventory; CU= Callous/Unemotional traits; IPPAr= Inventory of Parent and Peer Attachment revised; O=
Openness; C= Conscientiousness; E= Extraversion; A= Agreeableness; N= Neuroticism. *p<.05;**p<.01; *** p<.001.
Table 4.2 Linear regression analyses with YPI factors as outcome variables and attachment to parents as predictors
Interpersonal CU Lifestyle Variables B SE B β B SE B β B SE B β IPPAr mother -.15 .06 -.20** -.18 .05 -.28*** -.16 .05 -.25*** IPPAr father -.04 .05 -.07 -.01 .04 .74 -.05 .04 -.09 R2 .06 .08 .09 F for change in R2 7.38*** 10.27*** 11.51***
Note. CU= Callous/Unemotional traits; IPPAr= Inventory of Parent and Peer Attachment revised.
130 Table 4.3 Moderation analysis for mother and father attachment with YPI Interpersonal factor Interpersonal ∆R2 β Interpersonal ∆R2 β Step 1 .35*** Step 1 .33*** E .23*** E .22*** A -.55*** A -.57*** C -.13* Step 2 .38** Step 2 .34
IPPAr mother -.11* IPPAr father -.07
N .17** N -.12*
Step 3 .38 Step 3 .34
IPPAr mother x N .05 IPPAr father x N -.06
Step 1 .34*** Step 1 .31***
A -.55 A -.57***
C -.18** N -.18**
N -.23***
Step 2 .38** Step 2 .34**
IPPAr mother -.11* IPPAr father -.07
E .18** E .19**
Step 3 .38 Step 3 .34
IPPAr mother x E -.06 IPPAr father x E -.05
Step 1 .06*** Step 1 .01
C -.24*** E .12
N -.10 N -.02
E .11
Step 2 .37** Step 2 .34***
IPPAr mother -.11* IPPAr father -.07
A -.54*** A -.57***
Step 3 .36 Step 3 .34
IPPAr mother x A .01 IPPAr father x A .01
Step 1 .34*** N -.11 E .18** A -.59*** Step 2 .37** IPPAr mother -.11* C -.16** Step 3 .37 IPPAr mother x C .04
Note. N= Neuroticism; E= Extraversion; O= Openness; A= Agreeableness;
C= Conscientiousness; IPPAr mother= mother attachment; IPPAr father= father attachment.
131 Table 4.4 Moderation analysis for mother and father attachment with YPI
Callous/Unemotional factor CU ∆R2 β CU ∆R2 β Step 1 .31*** Step 1 .32*** E .03 E .03 A -.56*** A -.57*** C -.04 Step 2 .41*** Step 2 .38***
IPPAr mother -.16** IPPAr father -.04
N -.34*** N -.29***
Step 3 .41 Step 3 .38
IPPAr mother x N .07 IPPAr father x N -.04
Step 1 .39*** Step 1 .38***
A -.60*** A -.61***
C -.12* N -.25***
N -.29***
Step 2 .41** Step 2 .38
IPPAr mother -.16** IPPAr father -.04
E -.07 E -.07
Step 3 .41 Step 3 .38
IPPAr mother x E -.03 IPPAr father x E -.002
Step 1 .06*** Step 1 .03*
E -.15* E -.14^
C -.20** N -.20**
N -.27***
Step 2 .42*** Step 2 .38***
IPPAr mother -.16** IPPAr father -.04
A -.55*** A -.59***
Step 3 .42 Step 3 .38
IPPAr mother x A .01 IPPAr father x A -.02
Step 1 .38*** E -.08 A -.60*** N -.28*** Step 2 .41*** IPPAr mother -.16** C -.11 Step 3 .41 IPPAr mother x C -.02
Note. N= Neuroticism; E= Extraversion; O= Openness; A= Agreeableness;
C= Conscientiousness; IPPAr mother= mother attachment; IPPAr father= father attachment.
132 Table 4.5 Moderation analysis for mother and father attachment with YPI Lifestyle factor
Lifestyle ∆R2 β Lifestyle ∆R2 β Step 1 .43*** Step 1 .19*** E .33*** E .29*** A -.32*** A -.39*** C -.49*** Step 2 .45** Step 2 .22**
IPPAr mother -.17*** IPPAr father -.15*
N -.01 N .13*
Step 3 .45 Step 3 .26***
IPPAr mother x N -.02 IPPAr father x N -.20***
Step 1 .33*** Step 1 .11***
A -.28*** A -.34***
C -.50*** N .03
N -.12*
Step 2 .44*** Step 2 .22***
IPPAr mother -.17*** IPPAr father -.15*
E .34*** E .36***
Step 3 .44 Step 3 .22
IPPAr mother x E -.09 IPPAr father x E -.01
Step 1 .32*** Step 1 .08***
E .29*** E .31***
C -.52*** N .21**
N .04
Step 2 .44*** Step 2 .22***
IPPAr mother -.17*** IPPAr father -.15*
A -.27*** A -.33***
Step 3 .44 Step 3 .22
IPPAr mother x A .04 IPPAr father x A .03
Step 1 .21*** E .35*** A -.37*** N .16** Step 2 .44*** IPPAr mother -.17*** C -.46*** Step 3 .44 IPPAr mother x C .04
Note. N= Neuroticism; E= Extraversion; O= Openness; A= Agreeableness;
C= Conscientiousness; IPPAr mother= mother attachment; IPPAr father= father attachment.
133 Figure 4.1 Moderation effect of Neuroticism (N) on father attachment (with higher scores indication more secure attachment) and YPI Lifestyle
Lastly the relationship between attachment styles and YPI higher-order factors as well as FFM traits was investigated. As can be seen in Figure 4.2, the majority of the sample indicated having a secure attachment style (32%). There was also a large amount of individuals with anxious avoidant attachment (27%). Dismissing-avoidant attachment was the least indicated category (10%).
Correlation analyses were performed (see Table 4.6) between attachment styles, FFM traits and YPI higher-order factors. The secure attachment style was positively correlated with E, A and C, as was hypothesised. Secure attachment was also related to lower levels of N and higher CU traits, as measured by the YPI. Avoidant attachment was positively correlated with N and negatively correlated with E and A. Similarly, preoccupied attachment was also
significantly positively correlated with N and negatively with E.
0 0.5 1 1.5 2 2.5 1 2 3 4 5 YPI Li fe style Father attachment Low N Mean N High N
134 Figure 4.2 Attachment styles and percentage distribution
However, the correlation between preoccupied attachment and E was much weaker than for secure and avoidant attachment. Dismissing-avoidant attachment was significantly positively correlated to all psychopathic factors with the CU factor showing the strongest relationship. This attachment style was also negatively correlated with A.
Table 4.6 Spearman's rho correlations of normal personality traits, YPI factors and attachment styles
Factors Secure Avoidant Preoccupied Dismissing-Avoidant
YPI Interpersonal -.035 -.042 -.016 .201** YPI CU -.189** -.009 -.093 .389** YPI Lifestyle -.087 -.020 -.036 .170** N -.354** .328** .412** -.097 E .337** -.336** -.178** -.082 O -.007 .043 -.064 .065 A .281** -.178** -.095 -.307** C .176** -.070 -.114 -.006
Note. YPI= Youth Psychopathy Inventory; CU= Callous/Unemotional;
N= Neuroticism; E= Extraversion; O= Openness; A= Agreeableness; C= Conscientiousness. **p<.01.
4.5. DISCUSSION
Following on from the previous studies that examined the link between normal and
psychopathic personality traits with regard to antisocial tendencies, the current study sought
32.1 26.9 17.2 9.7 RQ: attachment styles secure anxious avoidant preoccupied dismissing-avoidant
135 to investigate the link between psychopathic traits and attachment. We also examined whether normal personality traits as measured by the FFM of personality would affect the relationship between psychopathic traits and attachment. This preliminary analysis was done also to determine whether or not attachment has an impact on psychopathic traits in the normal population and could potentially be investigated in subsequent studies to
examine its involvement in emotion-processing deficits seen in psychopathic individuals. The results of the current study suggest that attachment, the FFM and psychopathic traits are associated in predictable ways. Therefore attachment would be a theoretically interesting variable to consider in future studies examining the FFM, psychopathy and emotional processing and pro-and antisocial behaviours.
4.5.1. Attachment to parents
First the link between attachment to parents and psychopathic traits was examined.
Individuals with higher scores on all three of the YPI factors (Interpersonal, CU and Lifestyle) had lower attachment security to both mothers and fathers. This is in line with Kosson et al. (2002) who found this relationship in incarcerated adolescent males. However, similarly to Gao and colleagues who found mother attachment was significantly related to Factor 1 and Factor 2 of psychopathy (Gao et al., 2010), in subsequent analyses mother attachment was predictive of the YPI higher-order factors related to both primary and secondary
psychopathy. Mother attachment also remained a significant predictor of the Interpersonal,
CU and Lifestyle factors when the FFM personality traits were accounted for. The findings
suggest that including FFM traits into the regression model improves the model fit. At the same time attachment to mothers has an impact on psychopathic traits, which persists after
136 inclusion of FFM traits into the models. Individuals with less secure attachment to mothers were more likely to have higher psychopathic traits.
With the inclusion of the FFM personality traits, father attachment was a significant predictor of Lifestyle only. This relationship was moderated by N, such that for individuals with higher levels of N, insecure attachment to fathers increased secondary psychopathic traits (Lifestyle). Similarly, Forth and Flight (2007), investigated the psychopathy-attachment link in male incarcerated adolescents using the same attachment measures as used in the current study, and found father attachment to be significantly and negatively related to Factor 2 (Flight & Forth, 2007). These findings suggest that although father attachment may not play as strong a role for psychopathy as mother attachment does it is still to some degree influential, especially for individuals higher in Neuroticism.